Eigenvalues for factor analysis

The eigenvalue of 1st dimension reduced from 8.43 (explain %6.86 variance) to 6.28 (explain %5.12 variance).

2nd dimension increases its eigenvalues and variance explained. 3rd and 4th dimension remains the same.

Intepretation of dimensions

The top four dimensions can still be intepreted as before, with changes in top variable loadings on 1st and 4th dimension.

variable dim_1
sleep_eat_toilet 4.499382
sleep_dx 3.897170
med_cond_growth 3.223177
dev_motor 3.181388
feeding_dx 3.149210
enures 3.109189
dev_ld 2.693617
dev_id 2.678035
encopres 2.671462
gen_test 2.600246
variable dim_2
mood_or_anx 10.275140
mood_anx 7.161051
mood_dep 6.649120
mood_soc_anx 5.771604
age_at_eval_months 4.877607
mood_ocd 4.238542
behav_adhd 4.185975
attn_behav 3.966316
dev_lang_dis 3.445266
mood_sep_anx 3.088127
variable dim_3
med_cond_birth_def 12.249186
birth_def_bone 7.874764
birth_def_thorac 5.989751
birth_def_cns 5.645463
birth_def_thorac_heart 4.773771
birth_def_gastro 4.678376
birth_def_bone_miss 4.357501
birth_def_bone_spine 4.020091
birth_def_urogen 3.950929
birth_def_fac 3.931165
variable dim_4
dev_lang 4.613866
gen_test_fish_angel 4.547732
gen_test_fish_williams 4.331984
dev_lang_dis 4.190326
sleep_eat_toilet 4.120426
gen_test_mecp2 3.727107
gen_test 3.329941
gen_test_fish_digeorge 3.234268
gen_test_fish_oth 3.154239
gen_test_id 2.969846

First dimension are mostly sleep/eating/toileting problems, and several developmental disorders. The number of developmental disorder related variables are less than previous analysis.

Second dimension are mood related disorders such as anxiety, biopolar depression, etc. (Roughly the same as previous analysis)

Third dimensions are mostly birth defects. (Roughly the same as previous analysis)

Fourth dimensions are genetic testings (They are independent of birth defects/disorders).

K-means

Four-cluster is still meaningful.

K-mean results shows that the clusters are more separated on Mood&Depression rather than Development&ASD.